January, annual review. The head of procurement for a chain with 120 sites across the UK and Ireland puts the supplier scorecard on the table, four lines: delivery 78 points, quality 85, cost 90, service 70.

The supplier asks where the 78 comes from. She turns the page: seven of 31 deliveries were late. The supplier turns back: five of those were collected on the day promised after artwork approval. Late was the haulier, or the approval. Both are right. They are simply using two different clocks.

A supplier scorecard for packaging rarely fails on method and almost always on data. The scorecard is quickly built; the question is what goes into it and who measured it. We describe, from the side of the plant being rated, which figures we can report monthly, which we cannot, from which date delivery reliability counts, what separates a complaint from a phone call and how the annual review gets by with three numbers. This is not a standard: ISO 9001 requires supplier evaluation but does not prescribe the tool, and the GFSI-benchmarked schemes most UK retailers ask for expect packaging suppliers to be evaluated. How you rate is your business. With what, is the subject here.

What does a scorecard without supplier data measure?

It measures your goods-in and your haulier, not the supplier; the points are right, they just belong to someone else. A scorecard draws on three sources: goods-in with quantities and dates, quality with complaints, and buying with prices and response times. All three sit in your building. If the supplier contributes no data of its own, the scorecard rates a chain of events with the supplier at the end, without distinguishing which link broke. The seven late deliveries from the annual review are the example: five of them sat on the plant’s dock on the collection day. Without the supplier’s collection date they are seven delays; with it, two.

A sound supplier evaluation therefore starts with an agreement, not a spreadsheet: which figures does the supplier provide, which do you provide, and which only exist when both are put together? From our experience with chains and caterers the split is roughly half and half. The supplier knows production, collection and outgoing inspection; you know arrival, condition and consumption. Neither side alone has enough for a number that deserves to be called a rating.

OTIF: on time and in full, but counted from when?

OTIF counts a delivery only if it was on time and complete; on time means by the date promised from artwork approval. The wished-for date on the order does not count. The formula is simple: on-time and complete deliveries divided by all deliveries, times one hundred. Targets above 95 per cent are usual, above 98 often quoted. The formula is not the problem. The problem is the word “on time” in a trade where production only starts once the customer approves the artwork.

Our product page for cutlery sets gives the lead time as about 15 working days after artwork approval. Those last three words are the clock. An order received on 1 March whose artwork is approved on 20 March has a date in mid-April, not at the end of March. Whoever takes the order date as the start in the scorecard is rating their own marketing department too. For custom printed cutlery sets, whose sleeve carries the design, and for custom printed wet wipes, whose sachet is likewise printed only after approval, the same clock applies; both product pages give the 15 working days from artwork approval. The agreement that makes OTIF measurable therefore has three dates per order: approval date, promised collection day, actual collection day. The fourth date, arrival at your dock, belongs to the haulier, and under an Incoterm where you provide the transport, it belongs to you.

The “in full” has a trap of its own. Printed packaging is delivered in production runs, and a run may land a few per cent above or below the ordered quantity, because the press has start-up waste and the reel ends at some point. Whether 98 per cent of the quantity is “complete” must be agreed beforehand. A scorecard that counts every short delivery as incomplete without knowing the tolerance punishes a property of the process.

Complaint rate: against the specification, not against a feeling

A complaint is a measured deviation from a specification with a tolerance; what has no specification cannot be complained about, only discussed. The complaint rate is the share of deliveries with a documented quality problem out of all deliveries. Its meaning hangs entirely on what counts as a problem. “The sleeve feels thinner” is a phone call. “Grammage 84 instead of 90 at five per cent tolerance, six samples from three cartons, batch as on the delivery note” is a complaint. The difference is not pedantry but the precondition for the supplier being able to change anything: a number can be traced to a paper lot, a feeling cannot. Which lines a specification needs so that goods-in may measure is set out in our guide to packaging specifications with tolerances; the complaint rate is the figure that grows out of those lines.

Two things should be separated in the rate. The number of complaints and their severity: an open glue seam in one batch and a shade at the edge of tolerance are both a complaint, but not the same risk. And the time to an answer: a supplier that names a cause and proposes a measure within two working days is a different supplier from one that takes the pallet back after three weeks. Handling time per complaint is a figure the supplier can and should report itself.

Which figures can a packaging manufacturer report monthly?

Five figures it can report, because they arise in its building; two it cannot, because they arise in yours. The five we can give monthly start with the production date and batch number per delivery, from which you read traceability and the age of the goods. Then the promised and actual collection date per order, as the basis for OTIF. The measured values of the outgoing inspection per batch, meaning grammage, dimensions, colour difference against the limit sample, in the same units as your specification. The number of notified changes to material, coating or process, with the date of notification. And the handling time per complaint from receipt to stated cause. The two we cannot give: arrival time at your dock and the condition of the goods after transport. We do not see either, and a supplier that “includes” those figures has estimated them.

A hint from production on format: these five figures fit on one page a month, and they belong in a table, not in prose. A plant that cannot report them has no problem with you but with its own record-keeping. Ask at the first meeting; the answer says more about the supplier than any reference list. And the counter-argument we owe you: we are glad to report these figures, but the one that makes us look best, delivery reliability, hangs on your approval date. A customer who moves the artwork approval three times gets three new dates from us and an OTIF of one hundred per cent. Whether that helps them is another question.

Weighting: what counts more for printed packaging?

Delivery performance weighs more for printed packaging supplied to multi-site operators than in the average scorecard. A missing pallet leaves sites without packaging, and no wholesaler stocks the printed design. Common weightings put quality at 30 to 40 per cent, cost at 25 to 35, delivery performance at 20 to 25 and service at 10 to 20; a frequent scheme is 40, 30, 20, 10. That is a starting point, not a rule. For a caterer that calls off its napkins from a framework agreement with stock held at the manufacturer, delivery performance is almost a formality; for a chain without a central warehouse taking monthly deliveries it is half the business. Both may use the same scorecard, but not the same weighting.

A second difference from the average scorecard lies in price. For printed packaging the unit price is tied to quantity, version and Incoterm; a price change is often a quantity change or a paper-price clause, not the supplier’s decision. Whoever weights price at 30 per cent should have in that line which price is meant: the unit price per version at equal quantity, not the month’s invoice total.

The annual review: three numbers, one decision

Three numbers are enough for the annual review: OTIF from artwork approval, complaint rate against specification, handling time per complaint; the rest is conversation. With those three the decision the review demands can be taken: carry on as before, carry on with an action plan, or qualify a second supplier. OTIF below the agreed target and rising handling time is the pattern that justifies a second supplier; a high complaint rate with short handling time points more to a tight or unclear specification than to a poor plant. The supplier’s figures and yours lie side by side, and where they diverge is where the conversation takes place.

The head of procurement from January did not strike the 78. She added the column “promised collection day” to her table and set the approval date as the start. The next year it read 94, and two delays that really belonged to the plant were discussed for the first time instead of disappearing among seven.

KPI agreement before the first delivery

Eight points that should be in writing before the first delivery, so that the scorecard in January does not open with the question where the numbers come from.

  • Which date starts the lead time: artwork approval, not receipt of order.
  • Which quantity tolerance per production run counts as “in full”.
  • Which five figures the supplier reports monthly, in what format and by which day.
  • Which two figures your goods-in contributes and how they are linked to the batch number.
  • How a complaint is documented: measured value, nominal, tolerance, sample, batch.
  • What weighting the scorecard carries and why it is chosen that way for your operation.
  • Which figure triggers which consequence: action plan, follow-up audit, second supplier.
  • When the annual review takes place and who presents the figures on each side.

What you have read is the practice of a plant that gets rated; which KPIs your quality system prescribes and what your scorecard looks like is for you to set. But one experience holds everywhere: a supplier evaluation the supplier does not know about does not evaluate the supplier. It surprises them in January.

In short

  • OTIF is the share of on-time and complete deliveries out of all deliveries; targets above 95 per cent are usual, above 98 often quoted.
  • For printed packaging the lead time starts at artwork approval; delivery reliability is measured against the collection day promised from approval, not against the order date.
  • A complaint is a measured deviation from a specification with tolerance, documented with value, nominal, tolerance, sample and batch; without a specification there is no complaint rate.
  • A packaging manufacturer can report monthly the production date and batch, promised and actual collection day, outgoing inspection values, notified changes and handling time per complaint; arrival time and condition after transport arise at the customer.
  • Common scorecards weight quality 30 to 40, cost 25 to 35, delivery performance 20 to 25 and service 10 to 20 per cent; for multi-site operators without a central warehouse delivery performance weighs more.
  • ISO 9001 requires supplier evaluation without prescribing the tool; GFSI-benchmarked schemes expect packaging suppliers to be evaluated; the annual review gets by with OTIF, complaint rate and handling time per complaint.